Software Engineering Courses (SECourses)•3y ago

@Dr. Furkan Gözükara Hey... I don't remember... What CLIP-model did Stable Diffusion use? Default in

@Dr. Furkan Gözükara Hey... I don't remember... What CLIP-model did Stable Diffusion use? Default in CLIP Interrogator is VIT-L-14 and that's not it right?!

EyeSpyBekoAI•10/22/23, 8:02 PM

Runpod users, question for you. Do you know if it's possible to clone a pod, exactly as it is, and basically recreate an instance exactly with all the files, setup, etc., in it? For example, let's say I setup pod "X", downloaded Kohya, A1111, all extensions, files I need, etc. And then, let's pretend I come back to that pod tomorrow, and it has unavailable GPU. Is it possible to clone/recreate the whole setup, so that I can have a pod that has GPU available with my setup immediately setup?
Thanks!

KKallamamran @Dr. Furkan Gözükara Hey... I don't remember... What CLIP-model did Stable Diffu...

Furkan Gözükara SECourses•10/22/23, 8:10 PM

laion/CLIP-ViT-bigG-14-laion2B

EEyeSpyBekoAI Runpod users, question for you. Do you know if it's possible to clone a pod, exa...

Furkan Gözükara SECourses•10/22/23, 8:10 PM

shared disk space

Furkan Gözükara SECourses•10/22/23, 8:10 PM

or make your own template

EyeSpyBekoAI•10/22/23, 8:18 PM

Thanks for the response. I'll give those a shot

iamannakogan•10/22/23, 8:51 PM

Hi everyone! Good to be here!
A question: I am running kohya -> sdxl_training.py and got the following error:

Error(s) in loading state_dict for SdxlUNet2DConditionModel:
Missing key(s) in state_dict: "input_blocks.4.1.transformer_blocks.1.ff.net.0.proj.weight", "output_blocks.1.1.transformer_blocks.1.ff.net.0.proj.weight", "output_blocks.2.1.transformer_blocks.4.norm3.weight",
.... <many more lines>

I use the model from https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/resolve/main/sdxl_vae.safetensors. What could go wrong?

Iiamannakogan Hi everyone! Good to be here! A question: I am running kohya -> sdxl_training.py...

Furkan Gözükara SECourses•10/22/23, 9:17 PM

that is not a model

Furkan Gözükara SECourses•10/22/23, 9:17 PM

it is vae

Furkan Gözükara SECourses•10/22/23, 9:18 PM

1st : https://youtu.be/sBFGitIvD2A

YouTubeSECourses

Become A Master Of SDXL Training With Kohya SS LoRAs - Combine Powe...

In this tutorial, you will learn how to install Automatic1111 Web UI for SDXL. How to use LoRAs with Automatic1111 SD Web UI. How to install Kohya SS GUI scripts to do Stable Diffusion training. How to train LoRAs on SDXL model with least amount of VRAM using settings. All of the details, tips and tricks of Kohya trainings. How to do x/y/z plot ...

Furkan Gözükara SECourses•10/22/23, 9:18 PM

2nd : https://youtu.be/EEV8RPohsbw

YouTubeSECourses

How To Do Stable Diffusion XL (SDXL) DreamBooth Training (Full Fine...

Patreon post for config files : https://www.patreon.com/posts/89213064

FFurkan Gözükara SECourses 2nd : https://youtu.be/EEV8RPohsbw

iamannakogan•10/22/23, 9:20 PM

Thank you for the fast response!

FFurkan Gözükara SECourses laion/CLIP-ViT-bigG-14-laion2B

KallamamranOP•10/22/23, 10:15 PM

Hmmm can it fit 24GB VRAM?

KallamamranOP•10/22/23, 10:15 PM

In CLIP Interrogator I mean

KallamamranOP•10/22/23, 10:15 PM

The model is 10GB

KallamamranOP•10/22/23, 10:15 PM

Theroretically i could

FFurkan Gözükara SECourses sure here : https://youtu.be/kmT-z2lqEPQ

Deme•10/22/23, 10:48 PM

have been a couple of mounth , a simplified version ?
have you saw the pinokio.computer workflow ?
could you operate a guide using re_render and a fast workflow for most of 16vram users ? apreciated
I dont have money atm to subscribe to patreon i would otherwise for the re render script auto installer .

Marcelo•10/23/23, 2:06 AM

Does anyone have a comfyui workflow to load multiple loras?

FFurkan Gözükara SECourses laion/CLIP-ViT-bigG-14-laion2B

KallamamranOP•10/23/23, 4:58 AM

Holy hell.... That CLIP interrogation takes FOREVER

KallamamranOP•10/23/23, 6:34 AM

That took 8-12 minutes per image

Elisoli•10/23/23, 8:05 AM

yeah its really heavy , same here, (3060 12gb)

KallamamranOP•10/23/23, 8:15 AM

I really don't know if it was worth the time either

DR.Siva•10/23/23, 9:01 AM

when i do hires,fix on sdxl image (1.5 to 2 times) it ends with OutOfMemoryError: CUDA out of memory. ,anywy to fix it ?

KallamamranOP•10/23/23, 9:21 AM

Run the upscaling in Extras

LEGACIES•10/23/23, 10:16 AM

@Dr. Furkan Gözükara is it better to use the sd vae when training model with dreambooth?

KKallamamran Run the upscaling in Extras

DR.Siva•10/23/23, 11:28 AM

in extra , which model is best for SDXL ? and should i use both upscale ?

KallamamranOP•10/23/23, 11:38 AM

Trial and error

KallamamranOP•10/23/23, 11:39 AM

I like these the best, but it's amatter of taste and settings

KKallamamran Trial and error 😉

DR.Siva•10/23/23, 11:40 AM

Thanks a lot. I found this https://stable-diffusion-art.com/ai-upscaler/

Stable Diffusion ArtAndrew

How to use AI image upscaler to improve details - Stable Diffusion Art

AI Image upscaler like ESRGAN is an indispensable tool to improve the quality of images generated by Stable Diffusion. In fact, it is so commonly used that

DR.Siva•10/23/23, 11:41 AM

Taken from Reddit #Upscaler

DR.Siva•10/23/23, 11:43 AM

https://openmodeldb.info/ I think this is the place to download other Upscalers as well. CMIIW

OpenModelDB

OpenModelDB is a community driven database of AI Upscaling models. We aim to provide a better way to find and compare models than existing sources.

KallamamranOP•10/23/23, 1:06 PM

Anyone good at design training?! I mean... I can train a model of my face or an object, no problem. I can train a style, like a specific artstyle, but how do I train a design... Like... If I want to make an aeroplane in the Volvo designlanguage?! How would I train that?

KKallamamran Hmmm can it fit 24GB VRAM?

Furkan Gözükara SECourses•10/23/23, 1:15 PM

yes it fits

Furkan Gözükara SECourses•10/23/23, 1:15 PM

we run it as fp16- half precision

Furkan Gözükara SECourses•10/23/23, 1:15 PM

auto installer is on patreon : https://www.patreon.com/posts/sota-very-best-90744385

KallamamranOP•10/23/23, 1:16 PM

It fit, but man it took some time running a batch of images. About 8-12 minutes per image

MMarcelo Does anyone have a comfyui workflow to load multiple loras?

Furkan Gözükara SECourses•10/23/23, 1:16 PM

you can copy paste and append them

KKallamamran It fit, but man it took some time running a batch of images. About 8-12 minutes ...

Furkan Gözükara SECourses•10/23/23, 1:16 PM

much faster for me

Furkan Gözükara SECourses•10/23/23, 1:16 PM

https://www.youtube.com/watch?v=PNA9p94JmtY

YouTubeSECourses

Caption Images or Learn How To Prompt With Clip Vision of SDXL and ...

Download scripts from here : https://www.patreon.com/posts/sota-very-best-90744385

Full length Twitter video here : https://twitter.com/GozukaraFurkan/status/1712665890259447831

SOTA (The Very Best) Image Captioning Models Script For Stable Diffusion And More

1 Click install and use SOTA image captioning models on your computer. Supports 8 bi...

Furkan Gözükara SECourses•10/23/23, 1:17 PM

it is still slow though it is accurate

LLEGACIES @Dr. Furkan Gözükara is it better to use the sd vae when training model with dre...

Furkan Gözükara SECourses•10/23/23, 1:17 PM

i think can be used

DDR.Siva Taken from Reddit #Upscaler

Furkan Gözükara SECourses•10/23/23, 1:18 PM

nice

FFurkan Gözükara SECourses much faster for me

KallamamranOP•10/23/23, 1:20 PM

You're using VIT-L-14 in the video. Not laion/CLIP-ViT-bigG-14-laion2B

KallamamranOP•10/23/23, 1:20 PM

Right?!

KKallamamran Anyone good at design training?! I mean... I can train a model of my face or an ...

Furkan Gözükara SECourses•10/23/23, 1:21 PM

i think you need a good dataset and perhaps captions

KKallamamran You're using VIT-L-14 in the video. Not laion/CLIP-ViT-bigG-14-laion2B

Furkan Gözükara SECourses•10/23/23, 1:22 PM

KallamamranOP•10/23/23, 1:27 PM

The captions is easily created by CLIP interrogator and some manual labor. Question is what kind of dataset would be needed :/ All training videos (Including your fantastic videos) are of subject/object training or an art style. Not of a design

KallamamranOP•10/23/23, 1:27 PM

KallamamranOP•10/23/23, 1:27 PM

200 seconds in and is has barely started interrogating :/

KallamamranOP•10/23/23, 1:28 PM

It takes forever

KallamamranOP•10/23/23, 1:31 PM

I'm at 9% 430 seconds in

@Dr. Furkan Gözükara Hey... I don't remember... What CLIP-model did Stable Diffusion use? Default in

Similar Threads

Similar Threads

Similar Threads